You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I occasionally bump into the fact that the mincreshape *direction
options operate on image dimensions (the two fastest varying
dimensions, i.e., a slice) only. Things like this:
So I told mincreshape to force the step in x to be negative; but my
volume comes out unchanged. The explanation lives in the man page
under the +direction option:
+direction
Flip images to give positive step value for spatial
axes. Note that the flipping of spatial axes only applies to "image
dimensions". These are the two fastest varying (non-vector) dimensions
in the file. If you want to flip a non-image dimension, you can
convert it to an image dimension with -dimsize dimname=-1 (the -1
means don't really change the size). Check out the examples.
However, this is at best very confusing, if not just wrong. It seems
to me that a very specific instruction like -xdirection should "just
work" regardless of whether or not xspace happens to be an image
dimension. I have seen several posts over the years where people
scratch their heads over this (myself included :)
I can think of a few ways to address this:
modify mincreshape to "just work" for [+-][xyz]direction
throw a warning or error when [+-][xyz]direction is used while the
corresponding [xyz]space is not an image dimension
make this rather counterintuitive behaviour abundantly clear (e.g.,
description with every such option in the man page/help)
The text was updated successfully, but these errors were encountered:
I occasionally bump into the fact that the mincreshape *direction
options operate on image dimensions (the two fastest varying
dimensions, i.e., a slice) only. Things like this:
$ mincreshape -xdirection test.mnc test-x.mnc
Copying chunks:..................................................................................................................................................................................................................Done.
$ mincinfo test.mnc test-x.mnc
file: test.mnc
image: unsigned byte 0 to 165
image dimensions: xspace yspace zspace
dimension name length step start
-------------- ------ ---- -----
xspace 210 0.06 -6.33
yspace 274 0.06 -9.663
zspace 141 0.06 -4.8
file: test-x.mnc
image: unsigned byte 0 to 255
image dimensions: xspace yspace zspace
dimension name length step start
-------------- ------ ---- -----
xspace 210 0.06 -6.33
yspace 274 0.06 -9.663
zspace 141 0.06 -4.8
So I told mincreshape to force the step in x to be negative; but my
volume comes out unchanged. The explanation lives in the man page
under the +direction option:
+direction
Flip images to give positive step value for spatial
axes. Note that the flipping of spatial axes only applies to "image
dimensions". These are the two fastest varying (non-vector) dimensions
in the file. If you want to flip a non-image dimension, you can
convert it to an image dimension with -dimsize dimname=-1 (the -1
means don't really change the size). Check out the examples.
However, this is at best very confusing, if not just wrong. It seems
to me that a very specific instruction like -xdirection should "just
work" regardless of whether or not xspace happens to be an image
dimension. I have seen several posts over the years where people
scratch their heads over this (myself included :)
I can think of a few ways to address this:
corresponding [xyz]space is not an image dimension
description with every such option in the man page/help)
The text was updated successfully, but these errors were encountered: